dbench

Benchmark Kubernetes persistent disk volumes with fio: Read/write IOPS, bandwidth MB/s and latency. Based on https://github.com/leeliu/dbench

Usage

kind: PersistentVolumeClaim
apiVersion: v1
metadata:
  name: dbench
spec:
  storageClassName: ...
  accessModes:
    - ReadWriteOnce
  resources:
    requests:
      storage: 5Gi
---
apiVersion: batch/v1
kind: Job
metadata:
  name: dbench
spec:
  template:
    spec:
      containers:
      - name: dbench
        image: quay.io/pqsdev/dbench:main
        imagePullPolicy: IfNotPresent
        env:
          - name: DBENCH_MOUNTPOINT
            value: /data
          - name: FIO_SIZE
            value: 1G
        volumeMounts:
        - name: dbench-pv
          mountPath: /data
      restartPolicy: Never
      volumes:
      - name: dbench-pv
        persistentVolumeClaim:
          claimName: dbench
  backoffLimit: 4

1. Edit the storageClassName to match your Kubernetes provider's Storage Class kubectl get storageclasses
2. Deploy Dbench using: kubectl apply -f dbench.yaml
3. Once deployed, the Dbench Job will:
   • provision a Persistent Volume of 1000Gi (default) using storageClassName: (default)
   • run a series of fio tests on the newly provisioned disk
   • currently there are 9 tests, 15s per test - total runtime is ~2.5 minutes
4. Follow benchmarking progress using: kubectl logs -f job/dbench (empty output means the Job not yet created, or storageClassName is invalid, see Troubleshooting below)
5. At the end of all tests, you'll see a summary that looks similar to this:

==================
= Dbench Summary =
==================
Random Read/Write IOPS: 75.7k/59.7k. BW: 523MiB/s / 500MiB/s
Average Latency (usec) Read/Write: 183.07/76.91
Sequential Read/Write: 536MiB/s / 512MiB/s
Mixed Random Read/Write IOPS: 43.1k/14.4k

6. Once the tests are finished, clean up using: kubectl delete -f dbench.yaml and that should deprovision the persistent disk and delete it to minimize storage billing.

Notes / Troubleshooting

• If the Persistent Volume Claim is stuck on Pending, it's likely you didn't specify a valid Storage Class. Double check using kubectl get storageclasses. Also check that the volume size of 1000Gi (default) is available for provisioning.
• It can take some time for a Persistent Volume to be Bound and the Kubernetes Dashboard UI will show the Dbench Job as red until the volume is finished provisioning.
• It's useful to test multiple disk sizes as most cloud providers price IOPS per GB provisioned. So a 4000Gi volume will perform better than a 1000Gi volume. Just edit the yaml, kubectl delete -f dbench.yaml and run kubectl apply -f dbench.yaml again after deprovision/delete completes.
• A list of all fio tests are in docker-entrypoint.sh.

Contributors

• Lee Liu (LogDNA)
• Alexis Turpin
• YUUDJ

License

• MIT